Date of Award
Master of Arts (MA)
Languages, Literatures, and Linguistics
agree, control, partial, semantics, split, syntax
Arts and Humanities
This thesis presents a new theory of control in which the control predicate establishes control between PRO and the controller by Agreeing with both of them and checking their semantics at a later point of the derivation. Each DP possesses its own index as a syntactic feature. As the predicate Probes for these two Goals, it keeps track of the indexes of the controllee and controller with the control calculus, a part of the narrow syntax. The calculus is sensitive to the semantics of these nominal phrases after transfer to LF. This derives subject control in (1), which Culicover & Jackendoff (2001) argue exists due to the lexical semantic role Source on the Agent theta-role of promise.
1) John promised Mary PRO to eat pizza.
If the set containing the index of the controller is a subset of the set containing the index of the controllee, the derivation does not crash; otherwise, it crashes. This approach provides a derivation for the phenomenon of split control in (2), which Landau (2013) points out no account to date has been able to solve, and derives partial control in (3), along with deriving many other control phenomena.
2) John promised Mary PRO to meet up at 6.
3) John tried PRO to meet up at 6.
A major consequence of this solution is that it rejects the autonomy of syntax as proposed by Chomsky (1957); syntax provides a basis for semantics, which is not included in the grammar. But this paper argues that control phenomena show the need for a unified syntax and semantics approach to control, where both play a part in deriving sentences.
Satik, Deniz, "Control as Multiple Agree" (2018). Theses - ALL. 222.